1.3.1 The Argument from Mind–Brain Correlations

Science has gradually revealed a strong dependence between the mind and the brain. A striking early indication of this was the case of Phineas Gage (1823–60), a railroad worker who suffered an accident in which a metal rod was pierced through his brain (see Figure 1), yet he miraculously survived. But the accident led to a radical change in his personality: before, he was described as balanced and well-liked, but afterwards as “gross, profane, coarse, and vulgar” and “no longer Gage” (Reference MacmillanMacmillan 2000). If someone’s personality can fundamentally change based on brain injury alone, it indicates a strong dependence of the mind on the brain.

Figure 1 Phineas Gage

(Source: Wikimedia Commons)

Since then, neuroscience has offered more and more evidence for such dependence. We have discovered how consciousness can be affected by all kinds of physical events and interventions including not only severe injuries such as Gage’s, but also brain tumors, syndromes (such as Alzheimer’s), surgery (e.g., lobotomy or callosectomy), electromagnetic stimulation (from electroshock therapy to transmagnetic stimulation), and drugs (from anesthetics to LSD and antidepressants). And with the help of brain scanning techniques such as PET and fMRI (see Figure 2) we have discovered detailed correlations between specific conscious states and brain states (e.g., seeing red may correlate with specific activity in the visual cortex, anxiety with an overactive amygdala, and so on).

Figure 2 Brain images obtained by PET scans

All this can be taken to support a type of correlation known as supervenience between mind and brain. If A supervenes on B, there can be no change in A without a change in B. Thus, if consciousness supervenes on the brain, there can be no change in consciousness without a corresponding brain change (there can, however, be a brain change without change in consciousness, because not everything in the brain makes a difference to consciousness).

What explains this supervenience? Physicalists (e.g., Reference SmartSmart 1959; Reference HillHill 1991; Reference Block and StalnakerBlock and Stalnaker 1999; Reference McLaughlin, Gillett and LoewerMcLaughlin 2001; Reference MelnykMelnyk 2003) have argued that the best explanation is that conscious states are constituted by the brain states they supervene on. The most natural alternative is a dualist explanation, according to which non-physical consciousness is connected to the brain by means of fundamental laws of nature. The physicalist explanation is simpler and more elegant than the dualist explanation in that it doesn’t require positing extra laws of nature, and allows us to regard mind and brain as one thing rather than two. Other non-physicalist explanations, such as idealist and dual-aspect monist explanations, would also have to be regarded as unnecessarily complicated or otherwise implausible (though proponents of this argument tend to not explicitly focus on these alternatives).

1.3.2 The Argument from Physical Causal Closure

The argument from mind–brain correlations focuses on how the physical brain affects consciousness. The argument from physical causal closure (Reference KimKim 1989; Reference Papineau, Gillett and LoewerPapineau 2001; Reference MelnykMelnyk 2003; Reference Montero, Papineau and ClarkMontero and Papineau 2016) focuses on how consciousness affects the brain and body, or the physical world in general.

We take for granted that our conscious states have causal effects on the brain and body. If I feel pain from my arm being in an awkward position, that will cause me to move it. If I enjoy the taste of an apple, that will make me take another bite. If I have an intention to go for a walk, that will cause me to do it. And so on.

Yet, if consciousness is non-physical, it’s hard to see how this is possible, in view of the principle of physical causal closure. According to this principle, every physical effect, i.e., event that has a cause, has a complete physical cause. That is, some physical events, such as quantum events, may be truly random and without a cause, but those physical events that do have a cause – and this seems to include all human behaviorFootnote ¹⁶ – have a complete explanation in terms of physical causes.

Consider the case of someone moving their arm away from a painful position. This would have a complete physical explanation in terms of the nerves in their arm being stimulated by the way the arm is positioned, which causes electrical signals to be sent through their nervous system up to the brain, where it triggers the release of various neurotransmitters between neurons, culminating in a new electrical signal being sent back through the nervous system down to the arm, where the electricity causes their muscles to contract in such a way that the arm moves. We can thus explain (at least very roughly, in the case of this example) how behavior is produced without the mention of any conscious states, such as the feeling of pain or the intention to move the arm.

Furthermore, the signaling between the neurons can be regarded as implementing a kind of computation, with the neurons firing or not firing corresponding roughly to transistors in a computer turning on and off and thus representing 1’s and 0’s. This can physically explain (also very roughly) how our behavior can arise intelligently, or in a way influenced by and coherent with information received from perception, memory and so on.

As mentioned earlier, before modern science, the functions and abilities of conscious beings, such as intelligence and purposeful behavior, looked difficult to explain in physical terms, and this was one reason many regarded consciousness as non-physical. If all our functions and behaviors have a complete physical explanation after all, as the principle of physical causal closure claims, this reason to hold that consciousness is non-physical is undermined. But one might still think consciousness is non-physical for other reasons (such as it just seeming different from the physical). The argument from physical causal closure, however, combines the principle with two further premises in order to rule out non-physical consciousness altogether.

The principle of physical causal closure, by itself, rules out interactionist dualism, the view that conscious states are non-physical and cause physical effects different from those that would result from physical causes alone. But given the principle, non-physical consciousness could still be regarded as having no physical effects at all, as per epiphenomenalism. On this view, non-physical consciousness is produced or affected by physical brain states, but causes no physical effects in return. Another possibility is that our behavior has mental causes in addition to their sufficient physical causes. That is, our behavior might be overdetermined by more than one cause, both a physical brain state (that would have been sufficient to cause the behavior on its own) and non-physical conscious state (that is nevertheless present to also cause the exact same behavior).

The argument therefore adds that both epiphenomenalism and overdetermination are unacceptable, and that the only way to avoid them, if the principle of physical causal closure is correct, is to assume that conscious states are physical. If so, they could cause physical behavior, in a non-overdetermining way, in virtue of being constituted by physical brain states that we know cause behavior. Thus, physicalism offers the only plausible way of securing mental causation, i.e., that mental states cause physical effects, given physical causal closure.

To sum up the argument:

1. 1. Physical causal closure: Every physical effect (i.e., event that has a cause) has a sufficient physical cause.

2. 2. Non-epiphenomenalism: Conscious states have physical effects (i.e., behavior).

3. 3. Non-overdetermination: The physical effects of conscious states (i.e., behavior) do not have more than one sufficient cause.

Therefore,

1. 4. Physicalism: Conscious states are physical.

Another way of understanding the argument is as claiming that there are four possible positions one may take on mental causation, interactionism (which implies violation of physical causal closure), epiphenomenalism, overdetermination, and physicalism, among which physicalism is the most plausible option. The positions can be illustrated as in Figure 3.

Figure 3 Positions on mental causation

Why should we believe in the premises of this argument, or that physicalism is indeed the most plausible option? Epiphenomenalism and overdetermination seem quite implausible to most people, among other things because it just seems obvious that consciousness causes behavior, and it would seem like a strange coincidence that consciousness should always cause the exact same effects as the brain states they are correlated with (these and other arguments against epiphenomenalism and overdetermination will be discussed in more detail in Section 2).

The principle of physical causal closure can be supported by scientific evidence. Many have seen it as supported by the law of conservation of energy (see Reference MonteroMontero 2006, p. 384), because it may seem that if anything non-physical affects the physical world, it would add energy to it, thus violating conservation. However, one might think non-physical causes can only transfer energy to the physical that it previously received from it. If so, non-physical causes would be compatible with conservation of energy (Reference Papineau, Gillett and LoewerPapineau 2001; Reference MonteroMontero 2006). On the other hand, the scientific evidence could be interpreted as supporting not only that energy in general is conserved, but that physical energy is (Reference KoksvikKoksvik 2007). This is more of a threat to non-physical causes. But it’s debatable whether the evidence should be interpreted in this stronger way.

A more definitive argument derives from physiology, neuroscience, and other sciences of the human body in particular (Reference Papineau, Gillett and LoewerPapineau 2001). These sciences have revealed no sign of non-physical forces operating in the brain and body, which are the places influence from non-physical consciousness would be expected to show up. On the contrary, they strongly indicate that every physical event in the brain and body can be completely explained in physical terms. For example, as outlined earlier, we can explain why someone moves away from a painful position in terms of purely physical mechanisms. Of course, we don’t yet have complete physical explanations of all physical behavior, because we haven’t yet been able to map out every physical detail in the brain and body, so it can’t be definitively ruled out that signs of non-physical influence will eventually show up. But from the fact that every physical event in the brain and body examined so far has seemed completely physically explicable, and our research has reached an advanced stage, we can conclude that most likely they all are.

1.3.3 The Argument from Previous Explanatory Successes

The argument from previous explanatory successes points to how most other once mysterious phenomena have one by one been explained in physical terms. For example, before the scientific revolution, stars and planets were regarded as consisting of a different type of matter than things on Earth (since celestial bodies moved in circles whereas things on Earth moved downward in straight lines), and diseases were explained in terms of evil spirits or divine punishment. And even long after the scientific revolution, life was explained in terms of non-physical, vital forces (élan vital). But all these phenomena were eventually explained in physical terms by means of Newtonian physics, Pasteur’s germ theory,Footnote ¹⁷ and Darwin’s theory of evolution combined with Francis and Crick’s discovery of DNA. In addition, as also noted by the arguments for physical causal closure discussed previously, many of the major functions associated with consciousness (such as information processing and production of intelligent behavior) have already been physically explained, at least to a large extent.

According to physicalists (e.g., Reference SmartSmart 1959; Reference MelnykMelnyk 2003), we can conclude from this that consciousness itself (or its phenomenal or non-functional aspects) will eventually be physically explained, too. The reason why consciousness is so late to be fully physically explained may be that the brain is just far more complex than any other known physical phenomena.

Note that this argument is different from the argument from physiology in support of the principle of physical causal closure, according to which all physical events discovered in the brain and body so far seem to have a physical explanation, therefore we should expect all physical events (including so far unexamined or not fully explained events) to have a physical explanation. The argument from physiology concludes that all physical events (including those we usually take to be explained by consciousness) have a physical explanation, whereas the argument from previous explanatory successes concludes that consciousness itself has a physical explanation.

1.4.1 The Knowledge Argument

The knowledge argument claims that knowledge of consciousness can’t be deduced from any physical knowledge; therefore, consciousness is not physical. The most discussed versions of this argument are due to Reference NagelNagel (1974) and Reference JacksonJackson (1982).Footnote ¹⁹

Nagel argues that if bats are phenomenally conscious (as seems plausible that they are), they would have experiences very different from ours, because they have a very different sensory system based on echolocating sonars (i.e., they generate high-pitched sounds and register how they are echoed back to them). But no matter how much we learn about the physical characteristics of the bat’s sensory system, brain, and so on, we can’t deduce what it’s like to be the bat. Phenomenal facts must therefore be different from physical facts (at least, according to Nagel, on any typical understanding of the physical, corresponding roughly to the physics-based definition discussed in Section 1.2).

A possible objection to this argument is that since human brains are not configured to have sonar experiences, it’s not configured to imagine them either. So, the argument only demonstrates a limit to our imagination, not a limit to what is in principle deducible. Jackson’s version of the argument avoids this objection, because it appeals to a kind of experience most humans are capable of having, namely the experience of seeing red.

Jackson’s version of the argument is based on a thought experiment about a brilliant scientist called Mary, who grows up in a room where everything is black and white and has therefore never seen any colors. But in the room, she has gained complete knowledge of every physical fact about colors and colors vision from black and white books. One day she is let out the room and sees a red rose for the first time. She exclaims: “Wow! So this is what it’s like to see red.” In other words, Mary learns a new, phenomenal fact. But since she already knew all the physical facts about red and other colors, this must be a non-physical fact.

1.4.2 The Conceivability Argument

The conceivability argument claims that it’s conceivable that consciousness comes apart from the physical; therefore, it’s also possible that they come apart and consciousness is not physical. The conceivability argument goes back to Descartes (Meditations, VI; Discourse on Method, IV), who argued that we can conceive of consciousness existing without the body, or any physical world at all (as the physical world could conceivably be an illusion created by an evil demon), and that we can therefore see that consciousness is distinct from the physical. This version of the argument can be referred to as the argument from disembodiment.

Modern versions of the conceivability argument focus on how the physical world is also conceivable without consciousness.Footnote ²⁰ One of the most discussed versions of this kind of argument is Chalmers’ zombie argument (Reference Chalmers1996, Reference Chalmers, McLaughlin and Walter2009) (see also Reference KirkKirk 1974 and Reference KripkeKripke 1980 for important precursors). Zombies are defined as beings who are identical to us in every physical respect, including both their external behavior and their internal composition and structure – they have the exact same brain states. But they lack phenomenal consciousness – there is nothing that it’s like to be them. Chalmers argues that (1) zombies are perfectly conceivable, (2) if zombies are conceivable, then they are metaphysically possible, and (3) if zombies are metaphysically possible then physicalism is false.

Metaphysical possibility is roughly equal to logical or conceptual possibility – that is, to not involving any contradiction in terms, as the concepts of “married bachelor” or “square circle” do – or possibility in principle.Footnote ²¹ It should not be confused with nomological possibility, which is possibility given the actual laws of nature. Clearly, conceivability does not imply nomological possibility, as we can conceive of a number of things that are not possible given the actual laws of nature, such as people levitating, faster than light travel, and so on. But conceivability may plausibly be regarded as a guide to metaphysical possibility.

More precisely, the zombie argument takes conceivability to imply metaphysical possibility, and correspondingly, inconceivability to imply metaphysical impossibility, only under certain conditions. Roughly, the link holds only when we are conceiving of the things involved in terms of their real nature (Reference Nida-Rümelin, Alter and WalterNida-Rümelin 2006; Reference GoffGoff 2017), as opposed to in terms of how they superficially appear, or in terms of other contingent (i.e., inessential or coincidental) or imprecise characteristics.Footnote ²²

For example, for someone who conceives of a triangle superficially or very imprecisely (because they have not studied geometry), it could be conceivable that its angles don’t sum to 180 degrees, even though this is metaphysically impossible (assuming Euclidian space). But for someone who knows and understands all the characteristics of triangles and thinks about them in terms of their real nature, it should be inconceivable that their angles don’t sum to 180 degrees, and it could be concluded from this that this is also metaphysically impossible. When it comes to consciousness, we also seem to know its real nature, as its nature consists in what it’s like for us and this we seem to have direct access to. Therefore, (in)conceivability should be a guide to (im)possibility for consciousness in the same way that it is for geometrical objects, for example (for sufficiently qualified geometricians).

In order to make the final step from the metaphysical possibility of zombies to the falsity of physicalism, the argument also presupposes that if two things A and B are identical or A is constituted by B, then it’s metaphysically impossible for B to exist without A (as influentially noted by Reference KripkeKripke (1980) for the case of identity, and the case of constitution is closely related). For example, if the statue David is constituted by a piece of marble in a certain shape, it’s metaphysically impossible (i.e., impossible regardless of what the laws of nature may be) for that piece of marble to exist in that shape without the statue David also existing.Footnote ²³ In the same way, if consciousness is constituted by a physical state – as physicalism claims – it should be metaphysically impossible for that physical state to exist without consciousness. But, according to the zombie argument, this is perfectly conceivable and hence metaphysically possible, meaning that physicalism is false.

1.4.3 The Explanatory Argument

The explanatory argument claims that consciousness can’t be explained in physical terms; therefore, consciousness is not physical. This kind of argument goes back to Gottfried Wilhelm Leibniz (Monadology, paragraph 17), who asks us to imagine a conscious system (i.e., one that “thinks, feels and has perception”) large enough for us to walk into and inspect (as though it were a mill – hence the argument is known as Leibniz’ mill). Leibniz claims that no matter how much we learned about the system and its parts, we could never find anything to “explain a perception.”

One might think modern neuroscience has improved upon this situation. However, as Chalmers has argued (Reference Chalmers1995, Reference Chalmers1996) (see also Reference LevineLevine 1983 and Reference StrawsonStrawson 1994 for similar considerations), modern neuroscience and other relevant sciences still aren’t able to fully explain consciousness, at least not by means of any of their standard methods.

The standard methods of neuroscience are sufficient to solve what Chalmers calls the easy problems of consciousness. The easy problems consist in explaining the functions associated with consciousness (such as the ability to register and process information about the environment, generating intelligent behavior, and so on). To explain a function, all one has to do is identify a mechanism able to perform or implement the function. And neuroscience has identified physical mechanisms behind a number of important functions of consciousness (such as the production of behavior and the capacity to process information, as already outlined above), and it seems plausible that all the functions of consciousness can eventually be fully explained in the same way.

But neuroscience has no method for solving what Chalmers calls the hard problem of consciousness. The hard problem consists in explaining why phenomenal consciousness accompanies any of these functions, or how phenomenal consciousness arises from any physical processes at all (i.e., why aren’t we all zombies, in the sense of the conceivability argument)? When it comes to phenomenal consciousness, it seems neuroscience can discover correlations between phenomenal states and conscious states – for example, that the feeling of depression is correlated with low serotonin, or seeing red with some activity in the visual cortex. It might also identify what all the physical correlates of any conscious state have in common – for example, that they are characterized by high amounts of integrated information, as per the Integrated Information Theory, or that they are connected to a global neuronal workspace, as per the Global Workspace Theory, which are two leading general theories of consciousness in current neuroscience (the former will be discussed more detail in Section 4). But neuroscience cannot explain why these correlations hold: why couldn’t we have, for example, low serotonin but no feeling of depression (but rather, say a feeling of happiness or an experience of red), or high integrated information but no consciousness at all?

According to Chalmers, there is a principled reason why neuroscience is limited to finding mere correlations rather than full explanations, namely that its standard methods consist in finding reductive explanations, that is, explanations of phenomena in terms of the underlying physical parts or processes that make them up. But the only phenomena that can be reductively explained are functions or structures: functions can be reductively explained by finding physical mechanisms that perform them (as when solving the easy problems), and structures can be reductively explained by identifying the parts that compose them and the relations between them (e.g., a crystal, which can be understood as a structure, can be explained by identifying the molecules that compose it and the chemical bonds that hold between them). Phenomenal consciousness, in contrast, is not a mere function or structure: it may have functions and structure (e.g., a visual experience may have the function of informing us about the environment, and the structure of containing a distribution of different colors), but there is more to it than this (namely, what it’s like). Therefore, it can never be reductively explained. This argument, which can be regarded as a sub-argument of the explanatory argument, is known as the structure and function argument.

At this point, physicalists might invoke the argument from mind–brain correlations, according to which the correlations between conscious states and brain states can be explained by simply assuming that conscious states are constituted by their correlated physical states. But according to the explanatory argument, one cannot simply assume that conscious states are constituted by physical states, it must be shown that they are. To offer a reductive explanation is just to show how consciousness is physically constituted, and this is what the explanatory argument claims is impossible.

Physicalists may also invoke the argument from previous explanatory successes of science, according to which other phenomena that once seemed impossible to physically explain eventually were. Chalmers responds to this argument that all phenomena that have previously been reductively explained have been functional or structural phenomena. For example, diseases can be defined (roughly) by the function of making people sick, and life has been scientifically defined as a set of functions including metabolism, growth, homeostasis, reproduction, and so on. They can thereby be reductively explained by appeal to mechanisms such as germs (for diseases), or genes and DNA (for life). That a number of functional or structural phenomena have been physically explained gives us no reason to expect that a non-functional and non-structural, and thus entirely different, phenomenon such as phenomenal consciousness can be physically explained as well.

Note that the explanatory argument only claims that phenomenal consciousness can’t be reductively explained, or explained in physical terms. Phenomenal consciousness could still be explained non-reductively, or in non-physical terms, but this would result in a non-physicalist theory.

1.4.4 The Epistemic Gap – and Physicalist Responses

The knowledge, conceivability, and explanatory argument have something important in common. They each point out a disconnect between our knowledge or concepts of phenomenal consciousness and our knowledge or concepts of the physical, and conclude from this that phenomenal consciousness itself is not physical. As Reference Chalmers, Stich and WarfieldChalmers (2003) sums it up, they each point out an epistemic gap – a gap within what we can know or conceive (“epistemic” means regarding knowledge) – and conclude from this that there is an ontological gap – a gap in reality or between what is (“ontological” means regarding what there is or what reality fundamentally consists of). The arguments could therefore be jointly referred to as the arguments from the epistemic gap.

Physicalists have offered numerous and varied responses to these arguments (for an overview of the main strategies, see Reference Chalmers, Stich and WarfieldChalmers 2003). In general, however, physicalists may seem primarily motivated, not by specific errors perceived in the arguments against physicalism (in fact, some physicalists admit to finding them fairly compelling), but rather by how the arguments for physicalism seem so strong. The argument from physical causal closure, especially, carries much weight (as we saw, the explanatory argument against physicalism may cast some doubt on both the argument from mind–brain correlations and the argument from previous successes). Many philosophers find it far more plausible that there is some error in the arguments against physicalism – even if we can’t be sure exactly what the error is – than that phenomenal consciousness is epiphenomenal, an overdeterminer, or violates physical causal closure, as would follow from rejecting physicalism according to this argument.

Non-physicalists therefore need to answer the arguments for physicalism, the argument from physical causal closure in particular, in addition to offering arguments against it. From the next section onward, we will see how each of the main non-physicalist theories do this. But first, let us consider some further arguments against physicalism based on consciousness’ features of intentionality and unity.

1.4.5 Arguments Based on Intentionality

Intentionality (recall, the way conscious states can be about things in the world) can also be considered a problem for physicalism. One of the most influential arguments to this effect is based on the rule-following problem first raised by Ludwig Wittgenstein. According to this problem, when we consider any sequence of items, such as “1, 3, 5, 7 …,” it will always be compatible with a number of different rules, such as “add 2” (if so, the next item will be 9) or “add 2 three times, then subtract 2 three times” (if so, the next item will be 5) or “add 2 three times, then repeat the same number forever” (if so, the next item will be 7). That is, we can never infer which rule is being followed from the output of the rule alone.

In the same way, it does not seem possible to infer the meaning of an intentional state, such as a thought, or an expression of it, such as a word, from studying its physical manifestations alone. Reference QuineQuine (1960) argues that we can’t tell from someone’s behavior of uttering “rabbit” only in the presence of rabbits that this word refers to rabbits rather than, for example, undetached rabbit parts, which will always be present at the same time. Reference KripkeKripke (1982) argues that no physical facts determine whether someone apparently performing addition is not actually performing quaddition, which, just like addition, outputs the sum when applied to numbers up to 57 (or some other arbitrary number), but when applied to higher numbers always outputs 5, or relatedly, whether by the word “plus” they mean plus (in the sense of addition) or quus (in the sense of quaddition).

One might sum this up by saying that there is an epistemic gap from physical facts to intentional facts about meaning, in that the latter can’t be deduced from or explained by physical facts. Some, including Quine and Kripke, conclude from this that there are no determinate facts about meaning (such as about whether by the word “plus” we mean plus or quus) – because they don’t think any mental facts could determine meaning either. But others hold that there clearly are facts about meaning, and if physical facts don’t determine (or leave an epistemic gap to) them, facts about meaning must be regarded as non-physical.Footnote ²⁴

A typical response to this problem is to argue that although meaning isn’t physically determined in any simple or obvious way, it could be physically determined in more subtle or complex ways. For example, a number of attempts have been made at naturalizing intentionality – where by “naturalizing” is meant roughly the same as accounting for in physical terms – in terms of the physical abilities of conscious beings to track objects or features in their environment (Reference MillikanMillikan 1984; Reference DretskeDretske 1988; Reference FodorFodor 1990), that many regard as promising.

In view of this, some non-physicalists accept that intentionality can be physically explained, but still deny that phenomenality can (e.g., Reference KimKim 2005).Footnote ²⁵ One reason to treat intentionality differently in this respect is that, as we have seen, phenomenality appears non-functional and non-structural and arguably only functional and structural phenomena can be physically explained. But intentionality could be regarded as purely structural, because it may seem to consist mainly in a special type of relation (the “aboutness” relation) between consciousness and things in the world. Hence, there is less of a principled obstacle to physically explaining it.

Other non-physicalists hold that while intentionality cannot be fully explained in physical terms it can still be fully explained in phenomenal terms, or alternatively, in a combination of phenomenal and physical terms, as per what is known as phenomenal intentionality theories (Reference Horgan, Tienson and ChalmersHorgan and Tienson 2002; Reference Strawson and StrawsonStrawson 2008a; Reference Kriegel and KriegelKriegel 2013; Reference Mendelovici and BourgetMendelovici and Bourget 2014). According to such theories, what it’s like to be a given person – and this may include distinctively intentional or cognitive phenomenal qualities, such as feelings of intention, will, understanding, and so on – fully determines their intentional states, or (on some versions) determines them together with physical facts about various relations to the environment.

If intentionality is constituted either by the physical alone (as per attempts at naturalizing intentionality), the phenomenal alone or a combination of the phenomenal and physical (as per phenomenal intentionality theories), it poses no distinct problem for physicalism. For it to pose a distinct problem, intentionality must itself appear fundamental, or constituted by neither the physical nor the phenomenal, and this view is less often defended.

1.4.6 Arguments Based on Unity

The way consciousness seems distinctively unified has also given rise to arguments against physicalism. Descartes made an argument of this sort (Meditations, VI), according to which the mind is indivisible, whereas all physical bodies are divisible; therefore, the mind is non-physical.

More recent arguments claim, somewhat similarly, that the unity of consciousness must be accounted for by a subject distinct from the experiences it’s having, and subjects are simple and unified, but nothing physical is simple and unified in the same way (or at least no complex physical systems – fundamental particles may be simple and unified, but the argument tends to assume that subjects of experience couldn’t be constituted by single particles). Therefore, subjects are not physical (this or similar arguments have been defended by, e.g., Reference LoweLowe 1996; Reference Nida-RümelinNida-Rümelin 2007; Reference Hasker, Koons and BealerHasker 2010; Reference Barnett, Koons and BealerBarnett 2010; Reference SwinburneSwinburne 2013). Since, according to this argument, unity is a feature of subjects that are distinct from their phenomenal experiences, unity poses a problem for physicalism distinct from the problem of phenomenality.

As already mentioned, however, many philosophers, non-physicalists included, hold that the unity of consciousness can be accounted for by relations amongst phenomenal qualities themselves, rather than by a distinct subject, as per the deflationary view of subjects (Section 1.1; see also Reference StrawsonStrawson 2008b; Reference DaintonDainton 2010 for defenses of different versions of the deflationary view). If the unity of consciousness is just a relation between phenomenal qualities, then it may fall within the same category, and be explained in the same way, as these phenomenal qualities, and unity would not pose a distinct problem for physicalism after all (for other criticisms of unity arguments, see Reference Bayne, Loose, Menuge and MorelandBayne 2018b).

A related argument is the argument from identity across time. According to this argument, the same conscious being can exist at different points in time: for example, you now are the same person as you yesterday or you five years ago, and will be the same person tomorrow or in five years. This relation is not a matter of degree: for any conscious being in the past or future, you are either absolutely identical to them or not at all: you cannot be partially identical to anyone. But the physical body (including the brain) is not like this: your body tomorrow, for example, will contain at least a few different atoms and have various other physical differences (such as a few new hairs on your head, slightly different neuron configurations, etc.), and so is only partially identical to your body now. If you are absolutely identical to yourself in the past or future, then, it has to be in virtue of something non-physical that stays exactly the same, such as being the same non-physical subject (Reference Nida-Rümelin, Koons and BealerNida-Rümelin 2009; Reference SwinburneSwinburne 2013). Typical objections to this argument, however, include that perhaps we shouldn’t really take ourselves to be absolutely identical over time (Reference ParfitParfit 1971), or that absolute identity can in fact be accounted for physically, or in terms of mental properties, rather than a mental substance (see Reference Olson, Stich and WarfieldOlson 2003).

In the following sections, we will consider how the main non-physicalist theories can accommodate each of the arguments for and against physicalism mentioned so far. But for some of the arguments, there are no significant differences between the theories. When it comes to unity arguments against physicalism, we will see that some non-physicalist theories (such as substance dualism, subjective idealism, and emergent panpsychism) may be better able to accommodate them, and can therefore be seen as having an advantage – but only if one rejects the deflationary view of subjects, which many non-physicalists accept. When it comes to the arguments against physicalism based on phenomenality and intentionality, all non-physicalist theories can be regarded as equally accommodating – since they all agree that phenomenality is non-physical, and intentionality is widely regarded as being constituted either by phenomenality, the physical, or a combination thereof (in addition, they would also seem compatible with positing intentionality as fundamental alongside phenomenality).

The most significant differences are found in how the theories can respond to the arguments for physicalism, that is, the arguments from mind–brain correlations, previous explanatory successes, and physical causal closure – the latter, as discussed, being especially important – as well as the problems and objections that apply to each non-physicalist theory specifically, so this will be the main focus of discussion.

2.3.1 Epiphenomenalist Dualism

Given epiphenomenalist dualism, the psychophysical laws connecting consciousness and the physical go in one direction only: there are laws according to which physical causes produce mental effects, but no laws according to which mental causes produce physical effects.

Epiphenomenalists generally admit that their view is unattractive in a number of ways, but argue that it’s still less unattractive than the alternatives. More specifically, they tend to take both physicalism and interactionism to be refuted, or at least rendered highly implausible, by the epistemic gap and physical causal closure respectively. The arguments against epiphenomenalism, in contrast, don’t refute it nor render it nearly as implausible – at best, they point out reasons to dislike it or ways in which it’s counterintuitive (i.e., goes against what we find natural to believe), which are not valid reasons to reject it (Reference ChalmersChalmers 1996, p. 160). In addition, epiphenomenalists must endorse some of the arguments against other non-physicalist views (to be discussed in Sections 3 and 4) as more powerful than the arguments against epiphenomenalism.

The most basic argument against epiphenomenalism is that phenomenal states clearly appear to cause physical actions. Epiphenomenalists can simply respond that not everything that appears true is actually true, and in view of the evidence for physical causal closure, we should regard this appearance as false. One might object that, in that case, we might as well deny that consciousness exists in the first place, because our evidence for this is mainly that it appears to exist. Epiphenomenalists can respond that consciousness appears more strongly to exist than it appears to produce physical effects, or that appearances regarding the existence of consciousness are more (perhaps even absolutely) trustworthy in view of the direct access we have to our own consciousness, but this direct access does not extend to what our consciousness may or may not cause.

A second argument claims that epiphenomenalism is far less elegant than both physicalism and other kinds of dualism, because it posits consciousness as an idle “nomological dangler” (Reference SmartSmart 1959), rather than something properly integrated into the physical world. However, if the more elegant theories are refuted (physicalism by the epistemic gap and interactionism by physical causal closure), this is a moot point.

A third argument claims that epiphenomenalism is pragmatically untenable, or renders our lives meaningless practically speaking. This sentiment is powerfully expressed by Jerry FodorFootnote ²⁶: “… if it isn’t literally true that my wanting is causally responsible for my reaching, and my itching is causally responsible for my scratching, and my believing is causally responsible for my saying …, if none of that is literally true, then practically everything I believe about anything is false and it’s the end of the world” (Reference Fodor1990, p. 156).

Fodor here notes the radical implications of epiphenomenalism for how we look at our own lives. If our mental states have no physical effects, then not only do we (understood as mental beings) lack free will, that is, the ability to freely cause physical actions – as philosophers routinely worry about – we actually lack any will at all, that is, the ability to cause physical actions either freely or not.Footnote ²⁷ This turns us into passive spectators of our own lives – like the audiences of a movie, rather than actors in the drama. This would deprive our lives of much, if not most, of the meaning we ordinarily take it to have.

It could be objected that epiphenomenalists need not take mental states in general to have no physical effects, because some mental states, such as intentional states, may be physical even though others, such as phenomenal states, are non-physical. However, as discussed, many non-physicalists take intentionality to be non-physical as well. Furthermore, if our intentions can produce actions, but our phenomenal states can’t produce intentions (e.g., the feeling of love can’t cause you to intend to hug someone) it still puts strong limitations on our agency.

Epiphenomenalists may still respond that perhaps our lives just aren’t as meaningful as we think, or at least not meaningful in the way we usually think. In other words, that this is one of the features of epiphenomenalism we might not like, but that doesn’t mean it’s not true.

A fourth argument against epiphenomenalism claims that it renders it inexplicable how consciousness could have evolved, because if epiphenomenalism is true, consciousness is inert and useless, and useless features don’t evolve (Reference Eccles and PopperEccles and Popper 1977; Reference PopperPopper 1978). In response, epiphenomenalists have noted that useless features may evolve as by-products of useful features, and given epiphenomenalism conscious states could be by-products of useful brain states (Reference BroadBroad 1925; Reference JacksonJackson 1982).

One might still wonder, however, why specific conscious states, such as pain and pleasure, have evolved as by-products of specific physical states and not other ones (Reference JamesJames 1890, pp. 143–144). For example, why has pain (or feelings of discomfort) evolved as a by-product of harmful physical states or processes, such as burning or suffocation, and pleasure (or feelings of comfort) as a by-product of beneficial states or processes, such as eating or breathing – rather than the other way around? In other words, why are conscious states such as pain and pleasure correlated with physical states that seem fitting or appropriate, in the way that avoidance-causing physical states seem to be fitting for pain, and attraction-causing physical states seem to be fitting for pleasure?

If epiphenomenalism is false, and pain causes avoidance behavior, it follows, for example, that creatures for whom harmful states (such as burning) are (fittingly) correlated with pain will avoid what’s harmful and thereby be selected for by evolution. Creatures for whom harmful states are correlated (unfittingly) with pleasure, on the other hand, will be attracted to and pursue what’s harmful to them, and thereby be selected against by evolution.

But if pain has no causal effects – as per epiphenomenalism – switching around pain and pleasure would make no difference to behavior and thus be ignored by natural selection. For example, a creature for whom harmful states cause pain might still pursue them and so be selected against, and a creature for whom harmful states cause pleasure might still avoid them and be selected for. To explain why (in most cases) pain is correlated with harmful states (as well as pleasure with beneficial states, for which the situation would be analogous), epiphenomenalism must be rejected – or so the argument goes.

In response to this argument, it has been argued that non-epiphenomenalist theories, such as interactionism and physicalism, actually face the same problem (Reference RobinsonRobinson 2007; Reference CorabiCorabi 2014). This is because epiphenomenalism can in fact explain the fitting correlations (assuming natural selection) by positing one-way psychophysical laws according to which pain is a by-product of avoidance-causing physical states in particular, and pleasure is a by-product of attraction-causing physical states in particular. Granted, this leads to a further explanatory question of why we have these particular one-way psychophysical laws rather than other ones. But to explain the fitting correlations, interactionism must posit specific two-way psychophysical laws according to which pain causes avoidance and pleasure causes pursuit in particular, whereas physicalism must posit specific psychophysical constitution relations according to which pain is constituted by avoidance-causing physical states, and pleasure is constituted by attraction-causing physical states in particular. These two-way laws or constitution relations are just as much in need of explanation as epiphenomenalism’s one-way laws.Footnote ²⁸ If this is correct, epiphenomenalism has no unique explanatory disadvantage after all.

Epiphenomenalism can also be confronted with other correlations between conscious states and physical behavior that seem fitting or appropriate in a way the view arguably can’t explain. For example, not only is it a mystery, given epiphenomenalism, why pain happens to be a by-product of physical states that cause avoidance (rather than attraction) behavior, one might also wonder why the phenomenal experience of red is a by-product of physical states that dispose us to utter “I’m seeing red” (rather than “I’m seeing blue,” or to exhibit some completely different behavior such as jumping or dancing). Or, if intentional states are also regarded as non-physical, one might wonder why intentions to, for example, raise one’s arm, are by-products of physical states that tend to produce arm-raising (rather than, e.g., leg-raising or no action) (see Reference Cutter and CrummettCutter and Crummett [forthcoming] for further examples of such correlations, or instances of psychophysical harmony, as they call it). Epiphenomenalists can respond to these arguments in the same way as above: non-epiphenomenalist theories, such as physicalism and interactionism, can explain these correlations only by positing specific two-way psychophysical laws or identity/constitution relations that are just as unexpected and improbable as the one-way laws epiphenomenalists may posit to explain the same correlations.Footnote ²⁹

A related problem, known as the paradox of phenomenal judgment (Reference ChalmersChalmers 1996), is that if our judgments about our experiences (such as “I’m experiencing red”) are not caused by the experiences themselves, as epiphenomenalists maintain, these judgments will be unjustified, because it’s typically held that judgments are justified only if they are somehow causally influenced by what they are about (this is known as the causal theory of justification). This is a problem in and of itself, and also because epiphenomenalism is itself a judgment about our experiences (“all experiences are non-physical and epiphenomenal”). It seems to follow that epiphenomenalism itself lacks justification and the view can be regarded as self-undermining.

In response to this problem, Reference ChalmersChalmers has argued (1996, Reference Chalmers2010b) that our phenomenal judgments (understood as mental thoughts rather than physical utterances) are different from other sorts of judgments in that they need not be caused by what they are about. Rather, when we have a thought or make a judgment about an experience, the experience can be understood as a part or constituent of the thought or judgment. It follows that these judgments can be justified without any causal connections (i.e., it offers non-causal theory of justification). It can also be argued that other non-physicalist theories should adopt the same theory of phenomenal judgments, among other reasons because it may seem necessary to account for our direct and immediate access to our own consciousness, which most non-physicalists hold that we have. Any objections one may have to this theory would therefore be a problem not just for epiphenomenalism but for other non-physicalist theories as well.

The most serious problems for epiphenomenalism, may therefore be those based on conflict with appearances, inelegance, and pragmatic consequences – which may not decisively refute it, but still put it at a clear disadvantage if there are other views that can avoid them (without incurring other at least equally serious problems).

2.3.2 Overdetermination Dualism

According to overdetermination dualism, the psychophysical laws go in both directions: physical states produce conscious states, and conscious states in turn produce physical effects such as behavior. However, conscious states will produce the exact same physical effects as the brain states they are correlated with. Our physical behavior will thus have two causes, a conscious state and a brain state, where each cause would have been sufficient on its own – in the same way that someone’s death can be doubly caused by a shot to the head and a simultaneous shot to the heart, where each shot would have been deadly on its own.

The overdetermination will also need to be systematic, which is to say that our behavior always and without exception has a sufficient physical cause in addition to its mental cause, and the physical cause will never fail such that the mental cause gets a chance to cause the behavior on its own (i.e., there has to be some mechanism such that if the physical cause fails, then so does the mental cause), as this would break physical causal closure.

Overdetermination dualism thus shares with epiphenomenalism the advantage of being compatible with physical causal closure. It also shares the disadvantage of being not very elegant. But unlike epiphenomenalism, it allows the mental to affect the physical.

Still, few philosophers prefer overdetermination over epiphenomenalism. The main reason is that it seems completely inexplicable why conscious states would, systematically and without exception, cause the exact same effects as the physical states they are correlated with. There is no reason to expect the psychophysical laws (connecting conscious states to behavior) to mirror the physical laws (connecting brain states to behavior) in this way, and since there are so many different ways they could possibly diverge, it seems extremely implausible that they would mirror each other simply by coincidence.Footnote ³⁰ Overdetermination dualism itself is therefore generally regarded as just as implausible (though see Reference MillsMills 1996 for a defense).

2.3.3 Interactionist Dualism

Interactionist dualism posits two-way psychophysical laws, according to which consciousness and the physical world mutually influence each other, where the physical effects of mental causes have no sufficient physical cause and are thus not overdetermined (except perhaps on rare occasions by coincidence, as opposed to systematically and without exception, as per overdetermination dualism). Consciousness may either cause physical behavior all by itself, or together with physical causes in such a way that both contribute but neither are sufficient – in the same way a forest fire can be caused by both drought and a lightning strike, where if one were missing the fire would not occur.

By denying that physical behavior has sufficient physical causes, interactionism is straightforwardly incompatible with physical causal closure. Interactionists must therefore argue that the evidence for this principle isn’t as strong as we think, all things considered. Strategies for this can be divided into two types, those that invoke quantum mechanics and those that don’t. We will begin with the latter.

As mentioned, the most important argument for physical causal closure is the argument from physiology, according to which there is no evidence of non-physical forces influencing the brain and body, and we are getting closer and closer to a complete physical explanation of all processes in the brain and body. There are, however, philosophers and scientists who disagree, at least if the physical explanation is understood as one that invokes only the laws of microphysics (i.e., the physics of entities or properties at the smallest level or reality, such as particles).

Strong physical emergentism is the view that genuinely novel properties or behaviors arise in complex macrophysical objects such as the brain, which can’t be accounted for by the laws of microphysics alone, but must rather be accounted for by fundamental laws of higher-level sciences such as chemistry, biology, or neurology. Philosophers who defend such strong emergence within the physical, or at least regard it as an open possibility in view of the current state of evidence, include Reference CartwrightCartwright (1983), Reference Boogerd, Bruggeman, Richardson, Stephan and WesterhoffBoogerd et al. (2005), and Reference GillettGillett (2016) (see also Reference BroadBroad 1925 for a classic defense). Interactionists could take such strongly emergent physical properties or behaviors to be caused by consciousness – at least when occurring in the brain.

Physicalists could retort that this is only a speculative possibility with no clear evidence to support it (Reference McLaughlin, Beckermann, Flohr and KimMcLaughlin 1992). They may also argue that strongly emergent physical properties or behavior, if actual, could still be explicable in terms of laws of physical sciences such as biology and neurology. Even if these sciences aren’t part of microphysics (because their laws, given strong physical emergentism, fundamentally apply only at the macrolevel), they may count as a part of physics broadly speaking. Thus, strong emergence would not require explanation in non-physical terms (only in non-microphysical terms). Another issue is that most strong physical emergentists take strong physical emergence to occur in systems beyond the body and brain, such as molecules or cells (including non-human cells such as plant cells or bacteria). In view of this, dualists must either posit consciousness in all these systems – thereby approaching panpsychism – or explain why strong physical emergence has a non-physical, mental cause in the brain but a physical cause elsewhere (Reference MørchMørch 2014).

Now for the responses based on quantum mechanics. Quantum mechanics describes physical systems by means of a wave function, according to which systems exist in a superposition between different states. For example, if the system is a particle, the wave function will describe it as superposed between many different positions and many different velocities. When we make a measurement, we will always find the particle in a definite location or with a definite velocity.Footnote ³¹ This is known as the collapse of the wave function (i.e., the reduction of its many possibilities to one). The probability of finding the particle in a particular position or with a particular velocity upon measurement can be predicted from the wave function. But no matter how much information we gather – or at least local information about anything able to directly causally affect the particle, as per Bell’s theorem of no local hidden variables – we can never predict the particular position or velocity the particle will collapse into with certainty. The outcomes of measurements of quantum systems thereby seem fundamentally indeterministic. Indeterminism is the view that some events are not fixed or necessitated by prior causes, laws of nature, or anything else, while determinism is the opposite view that all events are thus fixed or necessitated.

There are different interpretations of quantum mechanics, some of which preserve determinism by positing non-local hidden variables or that collapse actually never happens (e.g., Bohm’s pilot-wave interpretation or Everett’s many-worlds interpretation). But quantum mechanics can also be interpreted in ways that accept indeterminism and collapse. Interactionists have taken this to suggest a causal role for non-physical consciousness, either as influencing the outcome of collapse by deciding between or narrowing down the possibilities contained in the wave function, or by causing collapse but without necessarily influencing the outcome.

The idea that consciousness may play the role of influencing the outcome of quantum collapse has been seized upon not only by interactionist dualists (e.g., Reference Eccles and PopperEccles and Popper 1977) but also by libertarians about free will (e.g., Reference Van InwagenVan Inwagen 1983; Reference KaneKane 1985; Reference BalaguerBalaguer 2009). Libertarianism is the view that humans have freedom to choose between otherwise genuinely undecided (i.e., indeterministic) possibilities. It should be noted that interactionism does not imply libertarianism,Footnote ³² as the psychophysical laws governing interaction between consciousness and the physical world could be deterministic (thus securing that conscious beings have will, but not necessarily free will, as distinguished earlier). But the kind of interactionism that takes consciousness to affect the physical by influencing the outcome of quantum collapse (rather than in some other way) seems to at least strongly suggest libertarianism, since collapse is indeterministic.

Is it possible that non-physical consciousness influences the outcome of collapse? There are two main obstacles to this idea. The first is that indeterminism at the microlevel, such as a particle having an undetermined position, does not imply indeterminism at the macrolevel, that is, at the level of behavior or its precursors, such as it being undetermined whether a neuron is firing or not (and as a result, whether some action that would be triggered by this firing is performed or not). This is because even though the wave function does not determine the properties of individual particles, it does determine the average or other statistical properties of large numbers of particles. For example, if the wave function assigns a 20 percent chance that an individual particle will be found in area X, then we can expect close to 20 percent of particles of the same type to be found in area X as we measure large numbers of them. Macrolevel states, such as whether a neuron is firing or an action is performed, will usually depend on average or other statistical properties of a large number of particles, and therefore be statistically determined.

Still, there are ways in which indeterminism could be possible at the macrolevel after all. One possibility is that the collapse of single particles or other tiny systems within the brain can be amplified so as to lead to large scale effects, in a way similarFootnote ³³ to the butterfly effect known from chaos theory (according to which a single butterfly flutter could be decisive in triggering a hurricane) (Reference Koch, Murphy, Ellis and O’ConnorKoch 2009; Reference Aaronson, Cooper and HodgesAaronson 2016; see also Reference JedlickaJedlicka 2017 for a survey of the evidence for and against this hypothesis).

Another possibility is that macrolevel states (in this context, at the molecular level or higher, which is roughly the minimum threshold for being able to decisively influence action without relying on the kind of amplification just discussed) within the brain could themselves be superposed. This would require that these states achieve quantum coherence, which is roughly to be in a superposed and internally entangled state that is not destroyed by interference from the environment. Entanglement is a relation between different particles or objects where collapse of one instantly implies collapse of the other(s), regardless of how far away from each other they are located, and where the outcomes are strictly correlated (e.g., if particle 1 collapses to “spin up,” particle 2, which may be far away, must instantly collapse to “spin down”). A coherent macrostate constituted by entangled microstates of numerous particles will therefore behave as one superposed state. It is standardly held, however, that the brain is too “warm, wet and noisy” (as it’s commonly summed up) to maintain coherence for meaningful amounts of time (Reference TegmarkTegmark 2000). But others believe this is possible, and some have proposed concrete mechanisms for how it may occur (Reference Beck and EcclesBeck and Eccles 1992; Reference Hameroff, Penrose, Poznanski, Tuszynski and FeinbergHameroff and Penrose 2016).Footnote ³⁴

If brain states rather tend to decohere, or become entangled with the external environment, this would result in superposed macrostates involving both the brain and parts of the environment, all of which would then have to collapse together. The parts of the environment may include physical objects as well as other people. If consciousness influences the outcome of collapses of such superpositions, it would have an instant and direct influence over physical objects and other people of a kind we – at least on the face of it – don’t seem to have.Footnote ³⁵ It also leads to questions such as: what happens if different people involved in the same superposition (which could potentially be many) try to collapse it in different incompatible ways? Perhaps these and other issues can be sorted out, but this is somewhat unclear, as this possibility seems to have been little explored.

The second obstacle, however, is that even if one could identify superpositions whose outcome is decisive for behavior, the idea that consciousness influences these outcomes seems incompatible with quantum mechanics itself (Reference PereboomPereboom 1995; Reference Montero, Papineau and ClarkMontero and Papineau 2016). This is because even though quantum mechanics does not fix the outcome of a collapse, it does fix the probability of each outcome. And if consciousness gets involved in selecting the outcome, it must result in different probabilities than those assigned by quantum mechanics (i.e., higher for the selected outcome and lower for the others). Or, it might be suggested that conscious selection could in principle result in the same probabilities, but this would involve a coincidence akin to that posited by overdetermination dualism. It might also be suggested that quantum mechanics might not really fix the probabilities in cases where consciousness is involved, that is, deny that quantum mechanics is universal, but this is a radical claim from the scientific point of view, for which there is currently no evidence.

The idea that consciousness causes collapse without influencing its outcomeFootnote ³⁶ avoids this second obstacle (but not necessarily the first one, of identifying some superposition whose outcome is decisive for behavior and it would be possible for consciousness to collapseFootnote ³⁷). As mentioned, quantum mechanics tells us that superpositions will collapse (or at least appear to, depending on the interpretation) upon measurement, but it says nothing about what measurement consists in or why it leads to collapse. Physicists Eugene Wigner and John von Neumann both suggestedFootnote ³⁸ that measurement fundamentally consists in observation by a conscious being, but they did not develop the idea in detail. It resonated strongly, however, with authors associated with the 70s New Age movement, who went on to promote it to the public, and this may have contributed to giving it a reputation as unscientific (Reference Chalmers, McQueen and GaoChalmers and McQueen 2022, p. 4). Reference Chalmers, McQueen and GaoDavid Chalmers and Kelvin McQueen (2022) have recently proposed a precise and scientifically grounded version of the idea (see also Reference StappStapp 1993).

According to this proposal, consciousness is superposition-resistant, meaning that it may go into superposition, but when it does, it will quickly collapse on its own. When a superposed physical system (such as a particle, or neuron) is measured, it becomes entangled with consciousness, and will therefore quickly collapse, too (on an earlier version of the proposal, consciousness can’t be superposed at all; however, this turns out to imply that consciousness can’t change, given what is known as the quantum Zeno effect (2022, p. 16) – hence they modify the proposal to allow for superposition of consciousness after all).

Chalmers and McQueen show that this view is empirically testable in principle, though currently not in practice. It should be noted that the view is compatible with both dualism and physicalism (on the physicalist version, it’s the physical basis of consciousness that is superposition-resistant, rather than consciousness). Therefore, if it were to be empirically confirmed, it would not thereby confirm dualism as well. But it would still confirm the possibility of a causal role for non-physical consciousness compatible with physics, and thereby weaken (though not refute) the evidence for physical causal closure.

There is a question of whether the view that consciousness causes collapse without influencing the outcome offers the right kind of causal role for consciousness. Firstly, one might think that, if measurement is the only way by which the mental can affect the physical, our agency might seem to reduce to a kind of passive observation. But on Chalmers and McQueen’s view, the kinds of conscious states that cause collapse need not always be measurements. On their view, consciousness primarily collapses physical states in the brain. Some of these states will be connected to perception and further entangled with measurement devices, and thus constitute measurements, but others may be connected to volition and behavior. Still, if agency consists in collapsing brain states, without influencing the outcome, its role will be limited to one of merely “rolling the dice” where the outcome is entirely random – which can also be regarded as fairly passive, or at least more so than we would like.

Another question is what it means for consciousness to exist in a superposed state. How could it possibly feel to have an experience superposed between, for example, blue (throughout the entire visual field) and orange? Or, would each state in the superposition be experienced in isolation, as though by parallel subjects? Many proponents of consciousness collapse theories, such as Wigner, have been explicitly motivated by how it doesn’t seem to make sense for consciousness to be superposed (and so were Chalmers and McQueen, at least to some extent, before they discovered the conflict with the Zeno effect) – so if the more developed versions of the idea end up having to posit it after all, it loses some of its motivation and instead encounters a problem.

Summing up, the prospects for interactionism largely depend on how things turn out empirically, mainly in quantum physics and the physics of strong emergence. But even if it turns out in interactionism’s favor, there are also philosophical challenges that must be resolved.

4.2.1 The Argument from the Intrinsic Nature of the Physical

So far, we have seen that dual-aspect monists make the following claims:

1. 1. Physical properties are all structural (or relational, dispositional).

2. 2. Structural properties have realizers with intrinsic (or non-relational categorical) properties.

3. 3. Phenomenal or protophenomenal properties are intrinsic.

This suggests, but doesn’t establish, that (proto)consciousness is the intrinsic realizer of physical structure, because there could be other kinds of intrinsic properties that can play this role, too. For example, one might think there are other intrinsic properties that are unknown to us, or that we can know or imagine other intrinsic properties besides consciousness.

The argument could be completed by adding two further claims. Firstly, that consciousness is the only intrinsic property we know, and that all other purported options (such as shape or physical colors) turn out to be relational on examination (e.g., shape is reducible to spatial relations, and physical colors are mere dispositions to cause phenomenal colors in observers, or similar). It is also arguably the only intrinsic property we can imagine. Panprotopsychists might add that protoconsciousness could be known or imagined on the basis of consciousness.

This still leaves the possibility of positing intrinsic properties that are completely unknowable and unimaginable. But why posit unknown properties when there are known ones, i.e., consciousness or protoconsciousness, able to do the job? In other areas of inquiry, we usually posit unknown properties in our theories only when there are no suitable known ones (e.g., in the case of dark matter), and arguably, the same standard should be applied in this case.

To complete the argument, then, it can be added that:

1. 4. Phenomenal or protophenomenal properties are the only intrinsic properties we know.

2. 5. We should not posit unknown properties if there are known alternatives.

This gives the conclusion:

1. 6. Physical properties have realizers with (proto)phenomenal properties.

4.2.2 The “Solving Two Problems at Once” Argument (or Dual-Aspect Monism and the Arguments for Physicalism)

The argument discussed in the previous subsection is not based on the problem of explaining consciousness, but only on the problem of explaining the intrinsic nature of the physical. Dual-aspect monism could be defended based on this kind of argument alone (Reference SeagerSeager 2006; Reference Adams, Zimmerman and van InwagenAdams 2007). But it is another argument, that is based on the problem of explaining consciousness, that is mainly responsible for the recent resurgence of interest in the view.

According to this argument, positing consciousness or protoconsciousness as the intrinsic nature of the physical also offers the best explanation of how consciousness fits into the physical world, because it avoids the main problems of both physicalism and dualism at once (Reference Alter and NagasawaAlter and Nagasawa 2012; Reference ChalmersChalmers 2013). The main problem of physicalism is the epistemic gap. Since dual-aspect monism regards consciousness as non-physical, it is compatible with the epistemic gap – just like dualism. The main problem of dualism, on the other hand, is the argument from physical causal closure. In response to this, dual-aspect monism claims that, as the realizers of physical structure, consciousness gets an explanatory role compatible with physical causal closure (Reference StoljarStoljar 2001; Reference Chalmers, Stich and WarfieldChalmers 2003, Reference Chalmers2013; Reference Alter and NagasawaAlter and Nagasawa 2012).

The This latter response requires some elaboration. The response starts from distinguishing two different versions of the principle of physical causal closure, a narrow and a broad version (Reference StoljarStoljar 2001; Reference ChalmersChalmers 2013). The narrow version claims that every physical effect has a purely physical cause, or cause with only physical (or physically constituted) properties. But according to dual-aspect monism, a purely physical cause would be a purely structural entity, and structures arguably cannot really exist without being realized by something with intrinsic properties – in the same way software cannot really exist without hardware. If purely physical causes cannot even exist, they would not be sufficient to cause anything, so the narrow principle must be false.

The broad version claims every physical effect has a sufficient cause whose structural properties are all physical. Or put another way, that the only causal structure needed to explain physical effects is physical causal structure, where physical causal structure could be understood as causal relations that fall under physical laws. According to dual-aspect monism, the broad principle is what the scientific evidence for physical causal closure really supports, since this evidence mainly consists in the fact that all causal relations required to explain physical events examined so far fall under physical laws.

But the broad principle is compatible with dual-aspect monism, because it says nothing about whether the causal relations or structure also have intrinsic realizers, such as phenomenal or protophenomenal realizers. And if structure requires intrinsic realizers in order to exist, phenomenal or protophenomenal realizers would not be epiphenomenal or overdetermining, but rather have an essential explanatory role. This explanatory role may be described as causal, since by enabling the existence of physical causes (proto)consciousness would clearly be relevant to causation, but since this relevance would be different than that of physical properties, the role could also be described as constitutive or explanatory in a broader sense. Either way, consciousness will play a significant, non-redundant role as the realizer of physical processes, including our own physical behavior.

The broad version of the principle still rules out interactionist dualism, because interactionist dualism claims that some physical events (i.e., our behavior) require explanation in term of causal relations that do not fall under physical laws, but rather under fundamental psychophysical laws. In other words, interactionist dualism posits additional causal structure that is not physical, and which, unlike the structure posited by epiphenomenalist or overdetermination dualism, would be required to explain some physical events (i.e., behavior). The response therefore supports dual-aspect monism only, as opposed to non-physicalism more generally.

Dual-aspect monism can also respond to the other main arguments for physicalism, though in less original ways. The response to the argument from mind–brain correlations is roughly the same as that of subjective idealism, namely that as a monistic theory dual-aspect monism is as parsimonious as physicalism.Footnote ⁵⁴ Second, its response to the argument from previous explanatory successes is mainly the same as that of dualism, namely that previously explained phenomena are functional or structural, whereas consciousness is not, and therefore one cannot generalize from the former to the latter.

In addition to the problems of physicalism and dualism, dual-aspect monism also avoids the problems of subjective idealism, since in taking the physical world as observer-independent and as having the exact structure described by physics, it can be regarded as a form of realism about the physical world – though this is more rarely emphasized (given the low popularity of subjective idealism compared to dualism and physicalism).Footnote ⁵⁵

4.2.3 The Argument from Non-emergence

In addition to the argument from the intrinsic nature of the physical and the argument from “solving two problems at once,” there are two additional arguments especially worth noting.Footnote ⁵⁶

The first is the argument from non-emergence. This argument claims that consciousness cannot emerge from anything purely physical, or from putting together physical entities, such as particles, in the right way. But our own consciousness seems to emerge from particles in the brain. It follows that these particles cannot be purely physical but must rather have been fundamentally conscious or protoconscious all along. A stronger version of the argument claims that consciousness cannot emerge from anything non-conscious, or from putting together non-conscious entities in the right way. It follows that fundamental particles must be conscious, rather than merely protoconscious.

The weak version of the argument is discussed but ultimately rejected by Reference NagelNagel (1979), even though he cannot say exactly how it goes wrong, because he regards the pan(proto)psychist conclusion as too implausible (he also briefly invokes a version of the combination problem, to be discussed below).Footnote ⁵⁷ The strong version of the argument is both defended and endorsed by Reference StrawsonStrawson (2006b).

Both Nagel and Strawson claim that consciousness cannot be constituted by the physical, in view of the epistemic gap or closely related considerations. If it emerges from the physical, it must therefore be by something akin to causal production, or in accordance with a dualist psychophysical law.Footnote ⁵⁸ Against this, they both invoke considerations similar to those figuring in the interaction problem (Section 2.2), according to which such causal relations between the mental and the physical would be unintelligible and hence impossible. Nagel similarly claims that it’s unintelligible how any physical process can necessitate consciousness (and that causation must involve necessitation, rather than effects merely “following” causes, as per the Humean view that, as noted, can also be invoked against the interaction problem). Strawson claims, also similarly, that the emergence of consciousness from the physical would be an instance of brute emergence, understood as emergence that is unintelligible in principle (or “unintelligible even to God,” as he puts it), because there is simply nothing about the physical – if understood as completely devoid of consciousness – in virtue of which consciousness could emerge.

Note that this argument supports panpsychism (given the strong version) or panprotopsychism (given the weak version) understood simply as the view that fundamental physical entities are conscious or protoconscious, but not necessarily the dual-aspect versions, according to which fundamental physical entities are conscious or protoconscious specifically because this is the intrinsic nature of their physical structure.Footnote ⁵⁹ But the argument is still compatible with dual-aspect monism (and this is also the version of panpsychism Strawson endorses).

4.2.4 The Argument from the Integrated Information Theory

Another argument for panpsychism (in this case, panprotopsychism is not included) derives from the Integrated Information Theory (IIT), a neuroscientific theory of consciousness developed by Giulio Tononi (later joined by Christof Koch and others) (Reference TononiTononi 2008; Reference Tononi, Albantakis and OizumiTononi, Albantakis, and Oizumi 2014).

The central claim of the theory is that consciousness is correlated with maximal integrated information, or maximal Φ (“phi”) for short, which is a structural property with a precise mathematical definition. In short, everything that has maximal integrated information is conscious, and the higher the integrated information the higher the level of consciousness.

Very roughly,Footnote ⁶⁰ information (as IIT defines it, which is quite different from how it is otherwise defined) is as a measure of the extent to which a system causally constrains its own past or future state (i.e., how much can you tell about the next and previous state of the system by looking only at the system itself, ignoring external influences?), and integration is as a measure of the extent to which this information depends on the causal interconnections between the system’s parts (i.e., by cutting the system in two, thus severing the connections between the two parts, how much information in the previous sense is lost?). Finally, a system has maximal integrated information or Φ if it has more integrated information than any overlapping system, that is, any of its own parts or any larger system it is itself a part of.

The brain contains very high levels of Φ, especially in those areas that (according to IIT’s proponents) appear necessary for consciousness. But small amounts of Φ can also be found at the level of fundamental physics, for example, in protons and neutrons (Reference KochKoch 2012). It follows that these particles have a small amount of consciousness – unless they are part of a larger system with even higher Φ (such as a brain, cell, or molecule) which would then be conscious instead. IIT thereby implies panpsychism, or at least something quite close.Footnote ⁶¹ As with the argument from non-emergence, this panpsychism need not be of the dual-aspect sort, but it can be.Footnote ⁶²

Still, IIT is a controversial theory, both in view of the empirical (i.e., experimental and observational) evidence and for various theoretical reasons (see, e.g., Reference AaronsonAaronson 2014; Reference BayneBayne 2018a). An argument for panpsychism based on it would therefore be hostage to the empirical evidence turning out in its favor, and perhaps some further clarification and defense of its theoretical foundations.

4.4.1 The Incredulous Stare

The perhaps most common objection to dual-aspect monism, especially the panpsychist version, is that it’s simply too implausible or counterintuitive to suppose that particles or other simple, non-living entities are conscious. This is known as the “incredulous stare” objection.

One response is to claim that as long as a theory is coherent and supported by good arguments, as dual-aspect monists take their view to be, then it doesn’t matter if it goes against our intuitions. Furthermore, its counterintuitiveness may not be universal. Many people also seem to find panpsychism natural to believe, especially, perhaps, in societies more influenced by Eastern philosophy and religion, in which panpsychist and idealist ideas are more common than in the West.

It might also seem that every theory of consciousness has strange and counterintuitive consequences when you really think about it (Reference SchwitzgebelSchwitzgebel 2014). For example, dualism may lead to epiphenomenalism, and according to some, physicalism, in reducing consciousness to mere functioning or physical structure, comes close to denying that phenomenal consciousness really exists (at least as we ordinarily think of it) (Reference FrankishFrankish 2016; Reference StrawsonStrawson 2018).Footnote ⁶⁵ In other words, when it comes to consciousness, we already know that “the truth must be strange” (as Russell once said about physical objects [Reference Russell1912; see also Reference StrawsonStrawson 2006a], but might as well have said about consciousness).

It should also be noted that panpsychism typically doesn’t take everything to have unified consciousness, or to be conscious as a whole. Tables, chairs, or rocks, for example, are typically not regarded as having their own unified consciousness, but rather as consisting of particles each with a separate, simple consciousness. Only in some things, such as the brain, does simple consciousness combine to form a more complex unified consciousness (or cosmic consciousness “decombine” to form a less complex one, given cosmopsychism).

To the extent that the objection is still taken seriously, however, panprotopsychism has some advantage over panpsychism, because attributing mere protoconsciousness to particles (or the cosmos) may seem at least somewhat more plausible than attributing full-blown consciousness.

4.4.2 Pure Dispositionalism and Ontic Structural Realism

Another, more serious objection accepts that the physical world is purely structural, but claims that this structure doesn’t require any intrinsic realizers. The position that all physical properties are purely structural or relational with no intrinsic realizers or relata is known as ontic structural realism (Reference Ladyman and RossLadyman and Ross 2007). A similar position is dispositionalism (Reference Shoemaker and van InwagenShoemaker 1980; Reference MumfordMumford 2004) (already mentioned in Section 3.4), which claims that all physical properties are dispositional, with no categorical grounds or aspects.

One possible response to this objection is to grant that structural or dispositional properties may not require intrinsic realizers or categorical grounds.Footnote ⁶⁶ But if we assume that they nevertheless have them, and that (proto)consciousness plays this role, then we can have a theory of consciousness that avoids the problems of physicalism and dualism. A problem with this response, however, is that if intrinsic realizers are merely optional, then putting (proto)phenomenal in this role would not seem to give them an essential explanatory role. They would rather come across as a redundant extra, similarly to (even though not quite as the same as) given epiphenomenalism or overdetermination. Dual-aspect monism might therefore not avoid the main problems of dualism after all, at least not as clearly.

A better strategy for dual-aspect monists might therefore be to argue that intrinsic realizers are required. One important argument to this effect claims that by eliminating intrinsic properties, as ontic structural realism does, we end up collapsing the distinction between the physical and the mathematical (Reference van Fraassenvan Fraassen 2006). That is, if the physical is purely structural, then there would be no difference between the physical world and a mathematical object with the same structure (which could always be constructed). The view that physical reality is at bottom mathematical is known as Pythagoreanism (after Pythagoras, who famously claimed that “all is number”), and strikes most people as obviously false. But how can we be sure? One argument is that the physical world is clearly concrete, in some sense, whereas mathematical objects are purely abstract; therefore, they have to be different. Another argument is that every mathematical structure exists in the abstract sense, whereas there is only one physical universe – or, even if we accept the possibility of a multiverse, the number of physical universes should be still smaller than the infinite number of mathematical structures.

In response to this kind of argument, some ontic structural realists have simply embraced Pythagoreanism (including the consequence that every mathematically possible universe physically exists) (e.g., Reference TegmarkTegmark 2008). A more common response, however, is to claim that physical relations have a non-mathematical aspect, such as being causal. Similarly, dispositionalists typically claim that dispositions are not merely abstract relations but are rather characterized by a kind of concrete power or energy (which may be intrinsic but still not phenomenal or protophenomenal).

The philosophical debate around these issues is complex, but the perhaps simplest way dual-aspect monists could respond in turn is to argue that the nature of any such causal aspect, power, or energy would be quite mysterious. We have already mentioned (in connection with dualism’s interaction problem, Section 2.2) Hume’s claim that causal relations are unintelligible and mysterious, which he based on the claim that we don’t really experience any power, energy, or other qualities of the sort we ordinarily take to characterize causal relations. If this is accepted, dual-aspect monists may again invoke the premise that we should avoid positing unknown and mysterious properties (such as causal aspects or powers) when there are known ones that could do the job (in this case, the job of distinguishing the physical from the mathematical, and intrinsic phenomenal or protophenomenal relata are the already known candidates).

Another response is that causal powers may not be entirely unknown and mysterious, but only because we experience them in mental contexts, such as our own experience of agency and motivation. To avoid mysterious properties, therefore, we don’t need to eliminate causal powers or similar qualities, we can instead regard them as mental or protomental. And this, of course, leads back to panpsychism or panprotopsychism after all (see Reference Mørch and SeagerMørch 2019a for an overview of philosophers who have offered this kind of argument, and Reference MørchMørch 2018 for a defense of the argument).

4.4.3 The Combination Problem

The combination problem is the problem of explaining how complex macroconsciousness, that is, the kind found in humans, animals and perhaps other macrophysical entities or systems, arises from putting together entities with simple microconsciousness, that is, the kind found in particles or other microphysical entities, or protoconsciousness. Or, given cosmopsychism, it would be the problem of how less complex macroconsciousness arises from more complex cosmic consciousness or protoconsciousness (this version of the problem is also known as the decombination problem, as coined by Reference Albahari and SeagerAlbahari 2019) but for now, let us consider the problem as it affects non-cosmic dual-aspect monism.

The combination problem gives rise to the what is widely regarded as the most serious objection to dual-aspect monism.Footnote ⁶⁷ According to this objection, explaining mental combination leads to problems for dual-aspect monism that are strongly analogous to those of physicalism and dualism. It thereby undermines the “solving two problems at once”-argument according to which dual-aspect monism completely avoids these problems, which can be considered the most important argument for the view.

As already mentioned, constitutive panpsychism takes macroconsciousness to be constituted by microconsciousness, or micro-entities possessing it, related in particular ways. However, it seems that we can conceive of microconscious entities related in any way we want, without the whole collection having any unified macroconsciousness. William James, who raised an early version of the problem, famously illustrated the point as follows:

In other words, there seems to be an epistemic gap from microconsciousness to macroconsciousness (Reference GoffGoff 2009). If an epistemic gap can undermine physicalism’s claim that consciousness is constituted by the physical, it should also undermine dual-aspect panpsychism’s claim that it’s constituted by microconsciousness.

Panprotopsychists who specify the nature of the protophenomenal (such as panqualityists) seem to have the same problem. Panprotopsychists who regard the nature of the protophenomenal as unknown could argue that it can’t be ruled out that if we knew this nature the epistemic gap to consciousness would be closable. But as already discussed, this kind of appeal to ignorance can be unconvincing.

Emergent panpsychism, on the other hand, takes microconsciousness to causally produce macroconsciousness. This view is compatible with the epistemic gap (because to regard micro- and macroconsciousness as distinct cause and effect is to admit an ontological gap between them), but runs into a problem with finding an explanatory role for macroconsciousness. This is roughly because it seems all the particles or other microphysical structure in the brain (or other systems where mental combination could take place) would already be realized by microconsciousness. When macroconsciousness is produced, no extra physical structure (that is not constituted by particles or other microphysical structure) seems to be produced along with it. So, there is no physical structure for emergent macroconsciousness to uniquely realize, and it would end up either epiphenomenal or as an overdeterminer – just as given dualism.Footnote ⁶⁸ Emergent panprotopsychism would also face the same problem (in this case, regardless of whether the protophenomenal is conceived as known or unknown).

One of the most important responses to the combination problem, on behalf of constitutive panpsychism (it could perhaps also be adapted to panprotopsychism, but here we set that aside) is the phenomenal bonding view, due to Reference Goff, Brüntrup and JaskollaGoff (2016). According to this view, relations should be understood as having their own intrinsic nature, separate from the intrinsic nature of their relata. Furthermore, we might suppose that there is some physical relation whose intrinsic nature consists in co-consciousness, which is simply the relation of being experienced together from a single point of view, or in other words, the relation by which two experiences or qualities (such as one experience of phenomenal red alone and another experience of phenomenal blue alone) may merge to form a single unified experience (such as of phenomenal red and blue experienced together). There is no epistemic gap from a number of microexperiences, each with their own quality, being related by a physical relation with this intrinsic nature and a unified macroexperience with the same qualities.

One important problemFootnote ⁶⁹ with this solution is that it’s hard to see which physical relation the phenomenal bonding relation could correspond to. According to Goff, it would have to be some fundamental physical relation, such as the spatial relation. But it would follow from this that all things that are spatially related form a unified macroconsciousness. In other words, not only would tables, chairs and rocks be conscious as a whole after all, so would any random collection of particles (since all particles are spatially related), including, for example, half a rock and some amount of air around it, the set of a human being and a cat, and so on. Even panpsychists tend to find this view, known as universalism about mental combination, too implausible (though see Reference RoelofsRoelofs 2019 for a defense).

Another response has been to turn to cosmopsychism. Reference GoffGoff (2017) argues that it’s easier to account for how macroconsciousness can be constituted by some part or aspect of a larger, cosmic mind, than how it can be constituted by a number of simpler minds. But others hold that such top-down “decombination” would be just as hard to explain as bottom-up combination.

On the emergent side, one of the most important responses is the fusion view (Reference SeagerSeager 2010, Reference Seager, Brüntrup and Jaskolla2016; Reference MørchMørch 2014, Reference Mørch2019c). According to this view, when micro- or protoconscious particles come together in the right way, their micro- or protoexperiences will fuse or merge into a single unified experience. The micro- or protoexperiences will be absorbed into the new whole, and afterwards no longer exist as individuals – similarly to how small drops of water can fuse to form a larger drop, after which the small drops no longer exist as individuals. Emergent macroconsciousness thereby replaces its micro- or protoconscious base. If macroconsciousness replaces its base, it can also take over its explanatory role as the sole realizer of microphysical structure (e.g., a particle in the brain, which used to be realized by a single micro- or protoexperience, will now be realized by a part of a unified macroexperience), thereby avoiding epiphenomenalism and overdetermination.

The main problem with the fusion view is that there doesn’t seem to be any physical fusion in the brain, because the brain seems constituted by particles that in no clear sense are absorbed or disappear into the brain as a whole. One response to this problem is to appeal to strong physical emergence, as already mentioned in Section 2.3.3. If genuinely novel physical properties or behaviors arise in macrophysical systems, this could be explained by these systems having a novel, fused intrinsic nature (Reference MørchMørch 2014, ch. 6) (such an explanation of the novel properties would also work better than a dualist explanation, for reasons mentioned in Section 2.3.3).

Another option is to appeal to the Integrated Information Theory. IIT can be naturally combined with the fusion view, in the sense that it follows from IIT’s claim that consciousness corresponds to maximal Φ that the macroconsciousness of a system (which would have maximal Φ) would replace the micro-consciousness of its constituents (which would have lower and non-maximal Φ). Furthermore, according to IIT, the brain and other macrophysical systems could have maximal Φ even with no strong physical emergence. Maximal Φ could therefore be a good candidate for the physical correlate of fusion (Reference MørchMørch 2019c).